oughly and hence leave more food behind (the giving-up
density, or GUD) in risky food patches than safe food
patches (Brown 1988; Brown et al. 1992). All else being
equal, a forager should have a higher GUD as predation
risk increases, a higher GUD as its energy reserves increase
(asset protection principle Clark 1994), and a higher GUD
as the marginal value of the resource declines.
Giving-up densities track a forager’s optimal behavior in
a risky environment. Most desert rodents exhibit higher
GUDs in the open than in shrub microhabitat (Kotler et al.
1991); voles tend to exhibit increasing GUDs as vegetation
cover declines (Jacob and Brown 2000); GUDs increase with
moonlight or artificial illumination (Brown et al. 1988; Kot-
ler et al. 1991); and GUDs increase with the presence of
predators (Brown et al. 1988). Giving-up densities show
that predation risk imposes an additional foraging cost on
rodents, and this cost can be substantial. For desert rodents
the cost of predation risk appears to be 3 to 8 times higher
than the metabolic cost of foraging (Brown et al. 1994b).
Rodents must balance trade-offs of food and safety
among their different activities. Lima and Bednekoff (1999)
observed that many foraging experiments use an on /off sys-
tem of predation risk. They modeled how antipredatory re-
sponses should be affected by previous experience with
predators, and with the temporal scale of changes in preda-
tion risk. According to their “predation risk allocation hy-
pothesis,” prey should trade off their feeding effort and vig-
ilance in relation to temporal variation of predation risk.
Feeding efforts should be lowest during short periods of
high risk that punctuate longer periods of low risk. Con-
versely, foraging efforts should be highest during short pe-
riods of low risk that punctuate longer periods of high risk.
In terms of the model (Lima and Bednekoff 1999), stud-
ies with boreal voles and desert rodents have yielded mixed
results. Koivisto and Pusenius (2003) found that field voles
(Microtus agrestis) in the laboratory responded as predicted
to temporal variation in the presence of a least weasel (Mus-
tela nivalis;fig. 28.3). Sundell et al. (2004) exposed free-
ranging bank voles in large enclosures to cages with weasels
or weasel odor to mimick short- versus long-term exposures
to predation risk. While increased perceived predation risk
reduced the voles’ use of feeding trays and increased their
GUDs, neither prior experience nor the temporal scale of
exposure influenced the bank voles’ GUDs or foraging
efforts.
In aviary studies with heteromyid and gerbilline rodents
(Brown et al. 1988; Kotler et al. 1988, 1991; Kotler 1997),
animals showed clear and striking changes in GUDs in re-
sponse to nightly changes in owl activity or artificial illu-
mination. If the owls remained in the aviary for long se-
quences of nights, the rodents eventually increased their
foraging efforts, reduced their GUDs, and suffered higher
mortality. This change in activity was most striking when
kangaroo rats were offered no shrub cover. When owls
were present, kangaroo rats became inactive; all activity oc-
curred on nights without owls. However, during a long
sequence of nights with owls, the kangaroo rats shifted
abruptly from no activity to uncharacteristically high levels
of activity (Kotler et al. 1988).Do rodents fit the risk allocation hypothesis?
Thus, the clearest results from the predation risk alloca-
tion hypothesis (Lima and Bednekoff’s 1999) show how en-
ergetic demands force individuals to increase foraging de-
spite known risks, provided the risky periods are persistent.
A valuable contribution of future research will be to de-
termine how prey develop fear responses from combining
prior experiences of predator activity with immediate ex-
posures to predation risk. It may be that the best survival
strategy is to adopt a continuous and rather high level of
alertness to predators, regardless of past experiences. For in-
stance, a bird like the willow tit (Parus montanus) seems to
maintain a certain state of alarm at all times (Haftorn 2000).
Alternatively, experimental results that seemingly con-
tradict the predation risk allocation hypothesis may have
been too short and missed the timescale over which forag-
ers cultivate an underlying behavioral pattern for balancing
energy gain and safety. In the study by Kotler et al. (1988),
the experiment lasted over several days with changes in pre-
dation risk occurring nightly. Such studies reveal that het-
eromyid rodents (Brown et al. 1988) and gerbils (Kotler
et al. 1991) make appropriate and nightly adjustments to332 Chapter Twenty-Eight
Figure 28.3 The least weasel (Mustela nivalis), the world’s smallest, but most
voracious, carnivore, is responsible for vole mortality and fear in northern boreal
habitats. Additionally, the small weasels provide an alternative food source for
larger predators, especially owls during winter. The white winter coat is an
important survival adaptation. Photo by Seppo Laakso.